Abstract

Statement of the Problem: Access to high-quality, representative, and unbiased medical data is an enormous and urgent roadblock to progress in medical science. Machine Learning and AI Algorithms are only effective and reliable if the datasets they use to train can be trusted, and while these technologies are already being used successfully in medical research, their true potential is being limited by the lack of data to run them on. Ethically sourcing the data, navigating legislation, privacy concerns, and educating patients regarding the importance of their data pose additional and not-insignificant challenges. Medical data is currently fragmented, and locked away in hospitals with no easy way of accessing it for researchers, particularly smaller startups without the financial means of big tech. For hospitals, the storage of data represents both a legal obligation as well as a financial burden. Depending on the country of residence of a patient, their medical data is either forcibly requisitioned by the government, or bought and sold on an open market - in both cases the consent of the patient is not sought and often it is done without even the patient’s knowledge.